Agentic AI and Testable code base

Scripts

# Verify environment
node -v                                   # node version
npm -v                                    # npm version
git --version                             # git version
claude --version                          # Claude Code CLI version
claude doctor                             # check Claude Code installation health
# usage/quota is only available via the `/usage` slash command inside a session,
# there is no `claude usage` shell command

npm test                                  # run vitest
npx vitest run packages/internal/__tests__/greet.test.ts  # run a single test file

npm run cli -- <flow-name> [json-input]   # run a flow by name via packages/cli/main.ts
npm run cli -- migration
npm run cli -- greet '{"name":"Ada"}'
npm run cli -- phone-activations '{"inputPath":"assets/test_big_file.csv","outputPath":"assets/test_big_file.out.csv"}'

Agentic workflow

Install the plugin first:

/plugin install superpowers@claude-plugins-official

flowchart TD
    NewFeature(["New feature request"]) --> Brainstorm["skills: /brainstorming
    clarify intent & requirements"]
    Brainstorm --> Plan["skills: /writing-plans
    draft a step-by-step plan"]
    Plan --> PlanReview{"Human reviews
    the plan"}
    PlanReview -- "changes requested" --> Plan
    PlanReview -- "approved" --> TDD1["skills: /test-driven-development
    write failing tests first"]
    TDD1 --> Execute["skills: /executing-plans
    implement the plan"]
    Execute --> Review

    Bug(["Bug or Regression"]) --> Debug["/systematic-debugging
    root-cause before fixing"]
    Debug --> TDD2["skills: /test-driven-development
    reproduce with a failing test"]
    TDD2 --> ImplFix["implement the fix"]
    ImplFix --> Review

    Review{"skills: /code-review
    requesting-code-review"}
    Review -- "issues found" --> Execute
    Review -- "approved" --> Finish["skills: /finishing-a-development-branch
    merge or open PR"]
    Finish --> Done(["Shipped"])

Loading